DETECTION AND IDENTIFICATION OF ELECTRICAL FAULTS USING RANDOM FOREST CLASSIFICATION

Abstract

Electric power generation and their transmission over electrical power grids and systems are an integral part of human development. It has led to the efficient and steady growth of economies and general human development. Electric power generation and its conveyance over transmission lines are however like every system of engineering prone to faults and errors. The development of machine learning systems has been very instrumental in the detection and classification of phenomena and scenarios in various fields. In this study, we propose the use of a machine learning technique known as random forest classification to carry out a process of electric fault detection and identification using an approach of the binary and multiclass classification process. Using adequate preprocessing and classification, the proposed method in this study achieved a binary classification of fault or no-fault classification of 99.6% accuracy, and a multiclassification of type of electric fault identification performance of 89.45% accuracy. The proposed method, tools, and analysis carried out in this study are presented in this paper comprehensively.